How much is enough? Minimum sampling intensity required to capture extant genetic diversity in ex situ seed collections: examples from the endangered plant Sibara filifolia (Brassicaceae)

Sibara filifolia (Brassicaceae), Santa Cruz Island Rock Cress, an endangered annual plant endemic to the California Channel Islands which was presumed extinct for nearly 50 years, is currently known from only a few small patches on San Clemente Island (SCl) and Santa Catalina Island (SCa). In addition to protecting extant populations, recovery efforts have included ex situ seed collections to hedge against stochastic population losses. We examined genetic diversity and population structure in the wild and the effect of sampling intensity on the genetic diversity of ex situ seed collections using 13 species-specific microsatellite loci. Much of the genetic variation on each island consisted of rare alleles; 33 % (SCl) or 56 % (SCa) of the non-fixed alleles occurred at frequencies ≤0.05. Patches on SCl were genetically depauperate (mean HO = 0.002) compared to patches on SCa (mean HO = 0.344). One patch on SCl was genetically distinct, while the remaining four patches were genetically indistinguishable. The three patches on SCa were genetically distinct from those on SCl, but not from one another. Simulated sampling for the ex situ seed collections found that 10 individuals was sufficient to capture 90 % of the diversity for normalized measures (NE, HO, and HE) in a depauperate (SCl) population, while 30 individuals was necessary in a diverse (SCa) population. However, sampling 125 (SCl) or 60 (SCa) individuals was necessary to capture 90 % of all the alleles present. These findings indicate that theoretical guidelines for ex situ seed collections that recommend targeting 50 individuals per population overestimate the sampling effort required to adequately preserve common alleles, but may underestimate the effort necessary to capture most alleles in wild populations.